messed up the title - it was supposed to say " please critique this " like the original thread.
it became obvious to me that i can completely eliminate one of 5 frequency bands from original design without much detriment so i did. here is the new design ( this is a horizontal cross section / view of array from the top ). the top of image is a waveguide that is forward facing towards the listener. the bottom of the image is a rear-firing subwoofer.
for reference here is the original thread when the design was 5 way, which is embarrassing now because of how much more complex and expensive it was for almost no benefit:
but that thread has a lot of information about what i am trying to accomplish with that design so i am including a link to it.
basically all the drivers remain the same except that 12" woofers are eliminated. previously they covered 60 hz to 150 hz range but now the 18" subs simply go up to 120 hz and the 5" midbass array stretches a bit lower to 120 vs 150 hz.
there was originally a purpose to 12" woofers a few design iterations back but they lost their usefulness over several redesigns so now they're gone.
it became obvious to me that i can completely eliminate one of 5 frequency bands from original design without much detriment so i did. here is the new design ( this is a horizontal cross section / view of array from the top ). the top of image is a waveguide that is forward facing towards the listener. the bottom of the image is a rear-firing subwoofer.
for reference here is the original thread when the design was 5 way, which is embarrassing now because of how much more complex and expensive it was for almost no benefit:
but that thread has a lot of information about what i am trying to accomplish with that design so i am including a link to it.
basically all the drivers remain the same except that 12" woofers are eliminated. previously they covered 60 hz to 150 hz range but now the 18" subs simply go up to 120 hz and the 5" midbass array stretches a bit lower to 120 vs 150 hz.
there was originally a purpose to 12" woofers a few design iterations back but they lost their usefulness over several redesigns so now they're gone.
Last edited:
referring to the color coded sketch the magenta square is a ribbon supertweeter: ( 3.5+ khz )
https://audioxpress.com/article/test-bench-fountek-neocd2-0-high-end-ribbon-tweeter
the green rectangle is a planar upper midrange ( 500 hz - 3.5 khz )
https://audioxpress.com/article/tes...bbon-transducer-from-radian-audio-engineering
orange is 5" midbass, of which 6" version was reviewed ( 120 hz - 500 hz )
https://audioxpress.com/article/test-bench-the-new-faital-pro-6-6pr160-midwoofer-midrange
and blue is 18" subwoofer ( 16 hz - 120 hz )
https://www.eighteensound.it/en/products/lf-driver/18-0/8/18ntlw5000
and white are the bass reflex ports ... tuned to 16 hz
https://audioxpress.com/article/test-bench-fountek-neocd2-0-high-end-ribbon-tweeter
the green rectangle is a planar upper midrange ( 500 hz - 3.5 khz )
https://audioxpress.com/article/tes...bbon-transducer-from-radian-audio-engineering
orange is 5" midbass, of which 6" version was reviewed ( 120 hz - 500 hz )
https://audioxpress.com/article/test-bench-the-new-faital-pro-6-6pr160-midwoofer-midrange
and blue is 18" subwoofer ( 16 hz - 120 hz )
https://www.eighteensound.it/en/products/lf-driver/18-0/8/18ntlw5000
and white are the bass reflex ports ... tuned to 16 hz
120 hz sub crossover chosen for two reasons:
1 - it is the cutoff frequency for LFE channel in movies
2 - it is the highest commonly used XO frequency in prosound and i am using prosound subwoofer drivers, so they can be guaranteed to work well to 120 hz
16 hz tuning is chosen because it is halfway between the 24 hz tuning used by JBL prosound subs using a similar 2269 driver and 10 hz tuning commonly used in home theater subwoofers. since this would also do duty for home theater but is a prosound driver 16 hz splits the difference and is also the frequency of the lowest pipe organ note.
now let's do some math:
20 khz / 3.5 khz = 5.7 X = 2.5 octaves ( here i am ignoring the fact that ribbon actually goes to 40+ khz )
3.5 khz / 500 hz = 7 X = 2.8 octaves
500 hz / 120 hz = 4.2 X = 2.1 octaves
120 hz / 20 hz = 6 X = 2.6 octaves ( here i am ignoring the fact that the system digs slightly under 20 hz )
the previous iterations of the design were bothering me because some drivers covered too wide a frequency range and others too narrow causing me to believe the design isn't optimized, and that feeling was correct. it wasn't.
i feel much better about this complement of frequency bands though. the 10 octaves of audible frequency range are now split almost evenly into four 2.5 octave chunks, which somehow just feels right.
it wasn't a primary objective to have even frequency bands but rather it was a sanity check to see if i am fully optimized. that is i would design for other things FIRST and then check whether i am balanced or not.
1 - it is the cutoff frequency for LFE channel in movies
2 - it is the highest commonly used XO frequency in prosound and i am using prosound subwoofer drivers, so they can be guaranteed to work well to 120 hz
16 hz tuning is chosen because it is halfway between the 24 hz tuning used by JBL prosound subs using a similar 2269 driver and 10 hz tuning commonly used in home theater subwoofers. since this would also do duty for home theater but is a prosound driver 16 hz splits the difference and is also the frequency of the lowest pipe organ note.
now let's do some math:
20 khz / 3.5 khz = 5.7 X = 2.5 octaves ( here i am ignoring the fact that ribbon actually goes to 40+ khz )
3.5 khz / 500 hz = 7 X = 2.8 octaves
500 hz / 120 hz = 4.2 X = 2.1 octaves
120 hz / 20 hz = 6 X = 2.6 octaves ( here i am ignoring the fact that the system digs slightly under 20 hz )
the previous iterations of the design were bothering me because some drivers covered too wide a frequency range and others too narrow causing me to believe the design isn't optimized, and that feeling was correct. it wasn't.
i feel much better about this complement of frequency bands though. the 10 octaves of audible frequency range are now split almost evenly into four 2.5 octave chunks, which somehow just feels right.
it wasn't a primary objective to have even frequency bands but rather it was a sanity check to see if i am fully optimized. that is i would design for other things FIRST and then check whether i am balanced or not.
as for physical construction there would be four subs per speaker, that would each be in their own cabinet that stack on top of each other like any prosound line array. each one would be around 23 inches high for a total of about 92 inches. that ensures both efficient use of plywood and that it fits under most common ceiling height of 8 feet.
i mean ceiling height and plywood size are related for a reason - both at 8 feet. by staying just under that you ensure that it fits under ceiling and also that your cuts don't run out of space on the plywood sheet. this is the beauty of such an array setup - efficient use of material and space.
each subwoofer would also have four 5" midbasses integrated in the front in a zig-zag fashion around the centerline - two on left of planar mid and two on the right. so a total of 4 x 4 = 16 midbasses per speaker.
planar midrange and ribbon supertweeters would NOT be integrated into these cabinets but installed AFTER the subwoofer / midbass units were stacked vertically and secured.
this is for many reasons:
1 - ribbons wouldn't run all the way to the top and bottom of the setup to save cost.
2 - ribbons would need to be stacked tightly without gaps and for that they must be able to span across two sub boxes.
3 - ribbons are expensive and fragile and i would rather remove them before transporting the sub boxes
i mean ceiling height and plywood size are related for a reason - both at 8 feet. by staying just under that you ensure that it fits under ceiling and also that your cuts don't run out of space on the plywood sheet. this is the beauty of such an array setup - efficient use of material and space.
each subwoofer would also have four 5" midbasses integrated in the front in a zig-zag fashion around the centerline - two on left of planar mid and two on the right. so a total of 4 x 4 = 16 midbasses per speaker.
planar midrange and ribbon supertweeters would NOT be integrated into these cabinets but installed AFTER the subwoofer / midbass units were stacked vertically and secured.
this is for many reasons:
1 - ribbons wouldn't run all the way to the top and bottom of the setup to save cost.
2 - ribbons would need to be stacked tightly without gaps and for that they must be able to span across two sub boxes.
3 - ribbons are expensive and fragile and i would rather remove them before transporting the sub boxes
minor tune-up.
Faital Pro 5PR160 is getting replaced with Faital Pro 6RS140
https://audioxpress.com/article/test-bench-faital-pro-6rs140-midwoofer-midrange
as you can see in the 120 hz to 500 hz target range distortion is basically nonexistent with this woofer. and that's measured in free air at 104 db !
this shows another benefit of something like a 4-way system. you can use just the portion of driver's bandwidth where it is clean. in this case it's a pretty narrow portion but that's all i needed anyway !
the downside versus 5PR160 is it's slightly less efficient ( due to rubber surround vs cloth accordion ), but it's worth for lower distortion. and besides, i didn't see any measurements for 5PR160 - only for 6PR160 and even if 5PR was as good as 6PR it would still not be as good as 6RS140 ...
i think it's that rubber surround that makes the difference in not creating distortion in that particular frequency range that i need ...
Faital Pro 5PR160 is getting replaced with Faital Pro 6RS140
https://audioxpress.com/article/test-bench-faital-pro-6rs140-midwoofer-midrange
as you can see in the 120 hz to 500 hz target range distortion is basically nonexistent with this woofer. and that's measured in free air at 104 db !
this shows another benefit of something like a 4-way system. you can use just the portion of driver's bandwidth where it is clean. in this case it's a pretty narrow portion but that's all i needed anyway !
the downside versus 5PR160 is it's slightly less efficient ( due to rubber surround vs cloth accordion ), but it's worth for lower distortion. and besides, i didn't see any measurements for 5PR160 - only for 6PR160 and even if 5PR was as good as 6PR it would still not be as good as 6RS140 ...
i think it's that rubber surround that makes the difference in not creating distortion in that particular frequency range that i need ...
i don't have measurements for 18NTLW5000 but i looked at measurements for other 18" subs from Eighteen Sound and they all more or less look like this:
which means that if crossed at 120 hz they are basically clean until you over-drive them ...
which means that if crossed at 120 hz they are basically clean until you over-drive them ...
minor tune up:
since i took a dump on Radian in another thread why not replace it with GRS version and save $2,000 in the process.
even though Radian version is more efficient and better built than GRS version i feel like it is not really brining anything new to the table over the GRS and can't justify being almost 3 times the price.
the GRS is well reviewed. i'm sure Radian is better but i think GRS will work too.
since i took a dump on Radian in another thread why not replace it with GRS version and save $2,000 in the process.
even though Radian version is more efficient and better built than GRS version i feel like it is not really brining anything new to the table over the GRS and can't justify being almost 3 times the price.
the GRS is well reviewed. i'm sure Radian is better but i think GRS will work too.
and while we're at it swap the Fountek for GRS "Hi-Res" which is either a clone or knock off but less than half the price ...
total parts list:
3.5 + khz - GRS RT3 Hi-Res - 24 per pair of speakers = $1,500
500 hz - 3.5 khz - GRS 10” planar, 16 per pair of speakers = $1,900
120 hz to 500 hz - 6” Faital 6RS140 - 24 per pair of speakers = $4,000
16 hz - 120 hz - 18” NTLW5000 ( vented at 16 hz ) - 8 per pair of speakers = $9,000
now that looks balanced ! as you go down in frequency every frequency band costs more to implement as it should because in music the lower you go in frequency the more energy there is and also the more displacement is required, thus more material and higher cost.
in prosound a subwoofer usually costs more than a speaker it works with and this is also the case here in our design here - sanity check passed.
actually insane how well priced those GRS ribbons and planars are.
now all you need is some used prosound amps from Ebay / Reverb, maybe fan mod them and you're in business.
total parts list:
3.5 + khz - GRS RT3 Hi-Res - 24 per pair of speakers = $1,500
500 hz - 3.5 khz - GRS 10” planar, 16 per pair of speakers = $1,900
120 hz to 500 hz - 6” Faital 6RS140 - 24 per pair of speakers = $4,000
16 hz - 120 hz - 18” NTLW5000 ( vented at 16 hz ) - 8 per pair of speakers = $9,000
now that looks balanced ! as you go down in frequency every frequency band costs more to implement as it should because in music the lower you go in frequency the more energy there is and also the more displacement is required, thus more material and higher cost.
in prosound a subwoofer usually costs more than a speaker it works with and this is also the case here in our design here - sanity check passed.
actually insane how well priced those GRS ribbons and planars are.
now all you need is some used prosound amps from Ebay / Reverb, maybe fan mod them and you're in business.
of course we first need @lowmass to confirm that GRS:
https://www.parts-express.com/GRS-RT3.0-8-Hi-Res-Neo-Ribbon-Tweeter-8-Ohm-272-206?quantity=1
is going to be as durable as Fountek:
https://www.madisoundspeakerstore.c...lk-5-ribbon-tweeter-rectangular-flange-black/
are they both sandwich diaphragm ? it seems GRS is just aluminum. so it means it is fragile ?
what justifies the price discrepancy between the two ribbons ?
https://www.parts-express.com/GRS-RT3.0-8-Hi-Res-Neo-Ribbon-Tweeter-8-Ohm-272-206?quantity=1
is going to be as durable as Fountek:
https://www.madisoundspeakerstore.c...lk-5-ribbon-tweeter-rectangular-flange-black/
are they both sandwich diaphragm ? it seems GRS is just aluminum. so it means it is fragile ?
what justifies the price discrepancy between the two ribbons ?
a few years back I bought same Fountek driver from parts express. Part of the reason I bought it was it said "sandwich" . at the time I was doing development work on 2 layer sandwich foil ribbons. I bought it to test and disect/ compare. Eventually I pull the ribbon and put under a scope to verify the so called sandwich construction. I could not at all see that it was anything but straight 12 micron Al foil. I also weigh it compared to 12 micron foil of same size and no difference. Im not sure what they mean by "sandwich" BUT I suspect it doesnt mean what we think it means. I replaced the so called sandwich foil with straight foil of same size and weight and sure enough the response and csd was identical.
In the sandwich protos I was developing at the time were to tame some higher frequency resonances. They did sound a bit sweeter and the csd showed a difference , HOWEVER the results were a more fragile ribbon. WHY? because it turns out that ya need a mass around what ya get with about 12 micron foil to optimize this drivers specs and that meant using two layers of 6 micron foil. Any durability increase ya get from sandwich is lost with the super tin foil. Also the thermal durability is lower with the two layers of 6 micron with visco in between.
I wouldnt be worried about using multiples above 3-4 khz on these drivers AND I believe the GRS is same unit
In the sandwich protos I was developing at the time were to tame some higher frequency resonances. They did sound a bit sweeter and the csd showed a difference , HOWEVER the results were a more fragile ribbon. WHY? because it turns out that ya need a mass around what ya get with about 12 micron foil to optimize this drivers specs and that meant using two layers of 6 micron foil. Any durability increase ya get from sandwich is lost with the super tin foil. Also the thermal durability is lower with the two layers of 6 micron with visco in between.
I wouldnt be worried about using multiples above 3-4 khz on these drivers AND I believe the GRS is same unit
the GRS looks like it's about 15 mm wide. i want 8 mm narrow.
what is the point of a wide ribbon - i am not trying to hit bass with it.
it seems a narrower gap would have higher flux thus better efficiency and maybe also better control of resonance as well as wider dispersion at higher frequencies.
seems like a win-win. but GRS of course caters to cheap skates who want to cross their ribbon straight to the subwoofer so they only offer the wider one it seems ?
what is the point of a wide ribbon - i am not trying to hit bass with it.
it seems a narrower gap would have higher flux thus better efficiency and maybe also better control of resonance as well as wider dispersion at higher frequencies.
seems like a win-win. but GRS of course caters to cheap skates who want to cross their ribbon straight to the subwoofer so they only offer the wider one it seems ?
narrow gap does not result in higher sensitivity. As the flux goes up, the surface area of ribbon goes down and its a wash sensitivity wise. It can have a more uniform flux across ribbon with narrower gap but thats all it does.
Wider ribbons do not have true uniform flux across ribbon. Its higher next to magnets than at center, but smaller gap can be more uniform
within practical limitations resonance control is mostly in the ribbons inherant mechanical properties dependent on material and shape.
Wider ribbons do not have true uniform flux across ribbon. Its higher next to magnets than at center, but smaller gap can be more uniform
within practical limitations resonance control is mostly in the ribbons inherant mechanical properties dependent on material and shape.
one thing is certain @lowmass and that is that nobody knows what is going on with these ribbons and buying them and testing them yourself is all anybody can do
even Vance Dickason is clueless - spewing a bunch of nonsense in his review of the Fountek - calling the plastic part "chamber" and that extruded aluminum part "a heatsink" ... embarrassing !
i wish people got banned for speaking out of their *** but instead negativity seems to be the only banable offence ...
i am still partial to narrower ribbons and i used my super eyeball to determine that GRS ribbon is 15 mm wide. i am like a geometry genius and stuff.
even Vance Dickason is clueless - spewing a bunch of nonsense in his review of the Fountek - calling the plastic part "chamber" and that extruded aluminum part "a heatsink" ... embarrassing !
i wish people got banned for speaking out of their *** but instead negativity seems to be the only banable offence ...
i am still partial to narrower ribbons and i used my super eyeball to determine that GRS ribbon is 15 mm wide. i am like a geometry genius and stuff.
I don't have the energy to digest all the posts, but here are some questions/potential issues:
1) You seemed to be assuming in the other thread that DSP was significantly influencing the horizontal dispersion of the JBL VTX V25-II. My speculation is that it's mostly about driver geometry, cross points, and the tweeter horn built into the midrange covers. You've deviated significantly from those design elements in multiple ways. Predicting the result is non-trivial.
2) As discussed in the other thread, I would model the array carefully. If that's not feasible, I would build one section and measure carefully. More points in image below. The more shallow the crossover and the higher the cross point, the more all of this matters.
3) The primary point I was getting at with the Danley reference in the other thread, and as summarized by my AI friend: "For drivers to combine coherently and form a uniform wavefront, their spacing must generally be around 1/4 wavelength or less for the frequency they are reproducing. This ensures that the drivers work together to create a smooth, coherent output rather than interfering with each other."
4) I would also build a Fome-Cor mock-up of the whole speaker and live with it in a room for a while (or a pair of them if you want the full dose). Speakers in real life tend to be a lot bigger than they are in your head or in a model. I think you said in another thread that your intention was that the speaker could be used in basically any room.
1) You seemed to be assuming in the other thread that DSP was significantly influencing the horizontal dispersion of the JBL VTX V25-II. My speculation is that it's mostly about driver geometry, cross points, and the tweeter horn built into the midrange covers. You've deviated significantly from those design elements in multiple ways. Predicting the result is non-trivial.
2) As discussed in the other thread, I would model the array carefully. If that's not feasible, I would build one section and measure carefully. More points in image below. The more shallow the crossover and the higher the cross point, the more all of this matters.
3) The primary point I was getting at with the Danley reference in the other thread, and as summarized by my AI friend: "For drivers to combine coherently and form a uniform wavefront, their spacing must generally be around 1/4 wavelength or less for the frequency they are reproducing. This ensures that the drivers work together to create a smooth, coherent output rather than interfering with each other."
4) I would also build a Fome-Cor mock-up of the whole speaker and live with it in a room for a while (or a pair of them if you want the full dose). Speakers in real life tend to be a lot bigger than they are in your head or in a model. I think you said in another thread that your intention was that the speaker could be used in basically any room.
that can get you into the ballpark but to get completely ruler flat beam width across all frequencies that literally does not budge from 100 degrees IMO you will need DSP trickery such as deliberately overlapping drivers to a calibrated extent at certain frequencies to narrow beam width.
it's the same as how good crossovers can get you decent FR but to get it RULER FLAT you need DSP.
keep in mind that DSP can not widen beam width - it can only be used to narrow it. same as how DSP cannot fill nulls in frequency response - it can only cut peaks.
well, i suppose you can slightly widen beam width by playing the center driver and flanking drivers out of phase just as you can fill nulls that aren't very deep but you get my point ...
get in the ball park ( but slightly wider than desired beamwidth ) in the physical domain then narrow it with DSP where it is too wide, and maybe in some extreme cases slightly widen it where it is too narrow ( if there is power reserve available to cancel out energy ).
in any case this is my speculation - i have never seen a write up about how this is done of if it is done at all. this is just what i think they are doing to get that response.
yes i am aware. this is why i desperately tried to find a 3" or smaller cone driver to flank the planar mid-tweeter but ultimately gave up seeing as they all have either high distortion ( like 20% ), low efficiency ( like 82 db ), or abysmal build quality ( wafer thin stamped baskets without so much as a rib on the flange for stiffness ).
3" cone would allow me to get the two banks of cones close together but instead i had to lower the XO point by going with a larger mid-tweeter planar capable of going lower.
let's do the math. current design calls for 10" planar model ( GRS and Radian are the same size, differing only in quality and price ) which has 3" wide radiating area. flanked by slightly angled 6.5" drivers each of them will contribute 3 inches to Center-to-Center spacing so a total of 3+3+3 = 9 inches center to center.
if 9" is 1/4 wavelength then wavelength is 36 inches and frequency is 375 hz ... i am just slightly pushing that with actual 500 hz crossover as the problem would only really manifest itself at 90 degrees off axis but my speaker isn't even going to have that wide a coverage based on a "V" shaped baffle that will be around 120 degrees wide, which means maximum off-axis angle of 60 degrees.
let's take it one by one ...
"how many wavelengths is this distance at cross point" ?
"what happens off axis when the drivers combine"
"what happens to planar output in this gap"
OK as i wrote elsewhere this is the one crossover that will be brick wall because it is the one XO where the two drivers will NOT be summing coherently, at least not off axis. so the answer is the drivers will not combine - they will operate at different frequencies with overlap too narrow to matter. this is one of those cases where it's something you can do with DSP that you can't do without.
now ADMITTEDLY the output of true ribbon supertweeter will diffract around the edges of its baffle and bounce off the main speaker ... this is why the main speaker will have to be lined with a thin foam to absorb HF wherever possible. i also expect that planar ribbon will not be a perfectly reflective surface but absorb some of that diffracted HF energy. perhaps 6" midbass units would have to be covered by plywood with a small hole in it ( as in Danley designs ) that way more foam can be put on the plywood. of course my design for this element would differ from Danley's as i believe his solution is crude. JBL has much more advanced "boundary integrators" in their arrays.
a "boundary integrator" in this context is like a cross between a phase plug and an acoustical low pass filter that also is bidirectional. in my case i might have to do that but also add Auralex foam on top of the surface to absorb diffracted back energy from the VHF ribbon.
as for what happens to planar output ... well the diagram shows 90 degrees angle but that was just easier for me to draw, actually it will be 120 degrees so the output of planar will easily get out of there. it will of course diffract around the VHF ribbon as well as reflect from it ... but i can put like a wedge shape diffuser or foam at the back of the VHF ribbon to help the sound go around it and also the ribbon face plate is 3 inches, which is the same width as the radiating area of the planar so per Olson ...
400 hz * 24" / 3" = 3200 hz.
so if we cross aroun 3.5 khz most of energy of the VHF ribbon will be directed forward ... but at the same time it won't be such a wide obstacle that energy from planar won't be able to go around it ... i mean not all will go around but enough will go around ... and because planar width is same as VHF ( very high frequency ) baffle width the same 3.5 khz crossover is also optimal for directivity of the planar ...
it wouldn't work if the ribbon faceplate was too wide for the crossover frequency but in this case it is just enough. 3.5 khz also is optimal XO frequency based on distortion charts of both ribbon and planar as measured by Vance Dickason. it is almost too perfect LOL.
"angle looks narrow"
yeah you're right. i just drew it that way because it's a hand sketch and my hand just felt like drawing that angle. it should actually be 120 degrees not 90 degrees.
with 90 degrees it would already begin to act a bit like a horn but i already have all the gain i need from the arraying and don't need additional horn gain
with a shallower angle i don't have to worry about any kind of cavity resonance or one driver's output reflecting off the opposite baffle etc ... so actually it was my intention even before drawing it to go with 120 degrees but i simply drew it the way my hand felt like drawing it ...
yeah you're right. i just drew it that way because it's a hand sketch and my hand just felt like drawing that angle. it should actually be 120 degrees not 90 degrees.
with 90 degrees it would already begin to act a bit like a horn but i already have all the gain i need from the arraying and don't need additional horn gain
with a shallower angle i don't have to worry about any kind of cavity resonance or one driver's output reflecting off the opposite baffle etc ... so actually it was my intention even before drawing it to go with 120 degrees but i simply drew it the way my hand felt like drawing it ...
"54 inches worth of arrows so maybe 57 inches wide"
again, this is just how my hand felt like drawing it. the actual size would be smaller.
actual size would be based on standard stairwell dimensions where to get a tall box through the turn in a home stairway it can't be bigger than about 30" x 40" however to be on the safe side i would keep it to about 24" x 32" ... but that's just the foot print. the height of the object can be anything but due to plywood dimensions the best height of the box would be 48" to avoid plywood waste.
so each array element would have to fit into a box with 24" x 32" x 48" dimension, but it would be navigated through the stairs ON THE SIDE so while moving 48" would be the height but when installed actually 24" would be the height, 32" the depth and 48" the width. four of those are stacked on top of each other to form the array with final dimensions of 32" X 48" X 96" ... actually it would be a bit shorter, around 92" to make sure it doesn't scrape the ceiling during assembly process ( if using 8 foot ceiling ).
with 48" width according to Olson ( the chart previously posted ) baffle step would be at 100 hz and my actual sub XO is 120 hz, however because the sub would be butted up against the wall that would actually reflect that wave and little to no output would be lost and in fact having it that close to the wall would avoid any kind of notch in response that can result from wall reflection coming out of phase ...
that's the reason for mounting the sub in the rear. if you but it up against wall or better yet room corner you get optimal loading for that sub without any out of phase cancelation, and at frequencies below 120 hz the sound would still be able to go around the 48" wide speaker.
again, this is just how my hand felt like drawing it. the actual size would be smaller.
actual size would be based on standard stairwell dimensions where to get a tall box through the turn in a home stairway it can't be bigger than about 30" x 40" however to be on the safe side i would keep it to about 24" x 32" ... but that's just the foot print. the height of the object can be anything but due to plywood dimensions the best height of the box would be 48" to avoid plywood waste.
so each array element would have to fit into a box with 24" x 32" x 48" dimension, but it would be navigated through the stairs ON THE SIDE so while moving 48" would be the height but when installed actually 24" would be the height, 32" the depth and 48" the width. four of those are stacked on top of each other to form the array with final dimensions of 32" X 48" X 96" ... actually it would be a bit shorter, around 92" to make sure it doesn't scrape the ceiling during assembly process ( if using 8 foot ceiling ).
with 48" width according to Olson ( the chart previously posted ) baffle step would be at 100 hz and my actual sub XO is 120 hz, however because the sub would be butted up against the wall that would actually reflect that wave and little to no output would be lost and in fact having it that close to the wall would avoid any kind of notch in response that can result from wall reflection coming out of phase ...
that's the reason for mounting the sub in the rear. if you but it up against wall or better yet room corner you get optimal loading for that sub without any out of phase cancelation, and at frequencies below 120 hz the sound would still be able to go around the 48" wide speaker.
i keep evaluating different ribbons and planars ( philosophically in my mind ) and readjusting driver complement but it doesn't affect crossover frequencies or design in any way. only the cost.
as of 5 minutes ago:
3.5 + khz - Aurum Cantus G3
66” array ( 6.6” X 10 )
2 X 10 @ 240 = $4,800
500 hz - 3.5 khz - Radian 10” planar
80” array ( 10” x 8 )
2 X 8 @ $340 = $5,500
120 hz to 500 hz - 6” Faital 6RS140
90” array ( 6.5” spaced at 7.5” in zig-zag in 4 boxes of 3 )
2 X 3 X 4 @ $170 = $4,100
16 hz - 120 hz - 18” NTLW5000
2 X 4 @ $1150 = $9,200
at some point i may post my thoughts on GRS vs Fountek vs Aurum Cantus for ribbons and GRS vs Radian for planars
for now the revelation i want to share is that apparently the 8" Radian uses Kapton but the 10" uses another plastic, probably Polyethylene Naphthalate ( PET ) because that's what GRS uses and Radian doesn't say what it uses.
my guess is PET is used for sound quality while kapton for power handling ( Kapton has significantly higher melting point than PET ). this way the smaller 8" still has about the same power handling as larger 10" but the response isn't as smooth. essentially the 10" is an audiophile midrange while the 8" is a prosound tweeter.
makes sense ?
cc: @lowmass
as of 5 minutes ago:
3.5 + khz - Aurum Cantus G3
66” array ( 6.6” X 10 )
2 X 10 @ 240 = $4,800
500 hz - 3.5 khz - Radian 10” planar
80” array ( 10” x 8 )
2 X 8 @ $340 = $5,500
120 hz to 500 hz - 6” Faital 6RS140
90” array ( 6.5” spaced at 7.5” in zig-zag in 4 boxes of 3 )
2 X 3 X 4 @ $170 = $4,100
16 hz - 120 hz - 18” NTLW5000
2 X 4 @ $1150 = $9,200
at some point i may post my thoughts on GRS vs Fountek vs Aurum Cantus for ribbons and GRS vs Radian for planars
for now the revelation i want to share is that apparently the 8" Radian uses Kapton but the 10" uses another plastic, probably Polyethylene Naphthalate ( PET ) because that's what GRS uses and Radian doesn't say what it uses.
my guess is PET is used for sound quality while kapton for power handling ( Kapton has significantly higher melting point than PET ). this way the smaller 8" still has about the same power handling as larger 10" but the response isn't as smooth. essentially the 10" is an audiophile midrange while the 8" is a prosound tweeter.
makes sense ?
cc: @lowmass